Enter opens the highlighted result. Esc closes.

Llama 3.1 70B Instruct vs Qwen3 30B A3B

Qwen3 30B A3B leads Llama 3.1 70B Instruct on the Artificial Analysis Intelligence Index, 7.6 to 6.6, at about the same output price ($0.50 vs $0.56 per million). Updated October 22, 2026; 43 benchmarks have a result for both.

Replace Llama 3.1 70B Instruct Replace Qwen3 30B A3B
Llama 3.1 70B InstructMeta Llama
Coding
No data
Overall
24.3
Output $/M
$0.56
0
Qwen3 30B A3BQwen
Coding
No data
Overall
38.2
Output $/M
$0.50
43
vs boards won of 43 shared

Llama 3.1 70B Instruct or Qwen3 30B A3B: which to pick

Neither model has a coding score yet, so the two cannot be ranked on coding.

Pick Llama 3.1 70B Instruct for a larger context window (131K against 41K tokens) and work measured by LiveCodeBench (Artificial Analysis run) or Terminal-Bench Hard (Artificial Analysis run), where it leads.

Model ALlama 3.1 70B InstructMeta LlamaNot benchmarkedNo coding benchmarkNo cost per solved taskModel BQwen3 30B A3BQwenNot benchmarkedNo coding benchmarkNo cost per solved task
Shared results
43
Llama 3.1 70B Instruct only
4
Qwen3 30B A3B only
9
Like-for-like categories
7 of 9

Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.

Which one for your work

Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.

At a glance

Coding index
Llama 3.1 70B Instruct not benchmarked, Qwen3 30B A3B not benchmarked.
Widest category
Math, 46.3 points apart on average over 5 shared benchmarks.
Widest single gap
AIME 2025 (Artificial Analysis run): Llama 3.1 70B Instruct 4.0% against Qwen3 30B A3B 72.3%, 68.3 points apart on a 0 to 100 scale.
Output price
Llama 3.1 70B Instruct $0.56 against Qwen3 30B A3B $0.50 per million output tokens, 1.1x apart.
Context window
Llama 3.1 70B Instruct takes 131K tokens, Qwen3 30B A3B 41K.

Benchmark by benchmark

First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.

Boards won, where both have a result
CategoryLlama 3.1 70B InstructBarsQwen3 30B A3BTies
Coding05
Intelligence02
Agentic work01
Reasoning06
Math05
Knowledge05
General019
All shared boards043
CodingLike-for-likeSix of these boards feed the Vibe Coding Index05
Coding benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
LiveCodeBench (Artificial Analysis run)23.2%50.6%Qwen3 30B A3B
Terminal-Bench Hard (Artificial Analysis run)3.0%6.8%Qwen3 30B A3B
WeirdML9.0%29.8%Qwen3 30B A3B
LMArena Text: coding12601337Qwen3 30B A3B
LMArena Text: software and IT12601340Qwen3 30B A3B
Neither model has a result on the other 34 boards here
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
Intelligence benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
AA Intelligence Index6.67.6Qwen3 30B A3B
Epoch Capabilities Index125.9137.3Qwen3 30B A3B
Agentic workLike-for-likePublic boards, shown as published, kept out of the index01
Agentic work benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
τ²-Bench Telecom15.2%26.0%Qwen3 30B A3B
BALROG27.9%Not measuredOnly Llama 3.1 70B Instruct measured
TheAgentCompany6.9%Not measuredOnly Llama 3.1 70B Instruct measured
Neither model has a result on the other 38 boards here
ReasoningLike-for-likePublic boards, shown as published, kept out of the index06
Reasoning benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
GPQA Diamond (Artificial Analysis run)40.9%61.6%Qwen3 30B A3B
Humanity's Last Exam (Artificial Analysis run)4.5%6.2%Qwen3 30B A3B
GPQA Diamond (Epoch AI run)44.2%61.7%Qwen3 30B A3B
LMArena Text: hard prompts12411314Qwen3 30B A3B
LMArena Text: expert12091314Qwen3 30B A3B
LMArena Text: hard prompts in English12611329Qwen3 30B A3B
Neither model has a result on the other 16 boards here
MathLike-for-likePublic boards, shown as published, kept out of the index05
Math benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
AA Math Index4.072.3Qwen3 30B A3B
AIME 2025 (Artificial Analysis run)4.0%72.3%Qwen3 30B A3B
OTIS Mock AIME3.6%62.8%Qwen3 30B A3B
MATH level 536.7%Not measuredOnly Llama 3.1 70B Instruct measured
LMArena Text: math12521355Qwen3 30B A3B
LMArena Text: mathematical work12531363Qwen3 30B A3B
Neither model has a result on the other 7 boards here
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index05
Knowledge benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
MMLU Pro (Artificial Analysis run)67.6%77.7%Qwen3 30B A3B
MMLU80.1%Not measuredOnly Llama 3.1 70B Instruct measured
LMArena Text: science12761321Qwen3 30B A3B
LMArena Text: medicine and healthcare12571318Qwen3 30B A3B
LMArena Text: legal and government12831314Qwen3 30B A3B
LMArena Text: business and finance12351309Qwen3 30B A3B
Neither model has a result on the other 12 boards here
Web designDirectional onlyPublic boards, shown as published, kept out of the index5 boards, none shared
Web design benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
Design Arena: websitesNot measured954Only Qwen3 30B A3B measured
Design Arena: UI componentsNot measured943Only Qwen3 30B A3B measured
Design Arena: game devNot measured898Only Qwen3 30B A3B measured
Design Arena: data vizNot measured967Only Qwen3 30B A3B measured
Design Arena: all code categoriesNot measured941Only Qwen3 30B A3B measured
Neither model has a result on the other 4 boards here
GeneralLike-for-likePublic boards, shown as published, kept out of the index019
General benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
IFBench34.4%41.5%Qwen3 30B A3B
Long context reasoning (AA-LCR)Not measured0.0%Only Qwen3 30B A3B measured
DTBench60.0%60.3%Qwen3 30B A3B
LMCA14.8%15.8%Qwen3 30B A3B
Fiction.LiveBenchNot measured40.6%Only Qwen3 30B A3B measured
Creative writing (Lech Mazur)Not measured75.3%Only Qwen3 30B A3B measured
LMArena Text12611317Qwen3 30B A3B
LMArena Text: instruction following12311283Qwen3 30B A3B
LMArena Text: creative writing12321271Qwen3 30B A3B
LMArena Text: English12941335Qwen3 30B A3B
LMArena Text: non-English12191295Qwen3 30B A3B
LMArena Text: Chinese12151347Qwen3 30B A3B
LMArena Text: German12221307Qwen3 30B A3B
LMArena Text: Japanese11321254Qwen3 30B A3B
LMArena Text: Korean11401261Qwen3 30B A3B
LMArena Text: French12621352Qwen3 30B A3B
LMArena Text: Spanish12531317Qwen3 30B A3B
LMArena Text: Russian12341291Qwen3 30B A3B
LMArena Text: PolishNot measured1299Only Qwen3 30B A3B measured
LMArena Text: multi-turn12561307Qwen3 30B A3B
LMArena Text: longer queries12411312Qwen3 30B A3B
LMArena Text: writing and language12361290Qwen3 30B A3B
LMArena Text: entertainment and media12301258Qwen3 30B A3B
Neither model has a result on the other 23 boards here
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
Image generation benchmarks for Llama 3.1 70B Instruct and Qwen3 30B A3B
BenchmarkLlama 3.1 70B InstructQwen3 30B A3BLeader
Neither model has a result on 5 boards here

What each workload costs

Three fixed token mixes at each model's listed price, so per-token rates become a bill.

Chat turn

1K input + 500 output tokens

Llama 3.1 70B Instruct$0.0008Fits in one request
Qwen3 30B A3B$0.0004Fits in one request

Qwen3 30B A3B costs less

Repository review

50K input + 3K output tokens

Llama 3.1 70B Instruct$0.030Fits in one request
Qwen3 30B A3B$0.0075Exceeds the context window

Qwen3 30B A3B costs less

Agent loop

200K input + 10K output tokens

Llama 3.1 70B Instruct$0.118Exceeds the context window
Qwen3 30B A3B$0.029Exceeds the context window

Qwen3 30B A3B costs less

Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.

Specification differences

The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.

Specifications of Llama 3.1 70B Instruct and Qwen3 30B A3B
SpecificationMeta Llama 3.1 70B InstructQwen3 30B A3B
MakerMeta LlamaQwen
WeightsOpen weightsOpen weights
ReleasedJuly 23, 2024April 28, 2025
Context window131K41K
Input $/Mtok$0.56$0.12
Output $/Mtok$0.56$0.50
Coding indexNot benchmarkedNot benchmarked
Independent measurements behind it00
Cost per solved taskNot computableNot computable
SourceOpenRouter listingOpenRouter listing

Questions

Is Llama 3.1 70B Instruct better than Qwen3 30B A3B for coding?

Neither model has a coding score yet, so the two cannot be ranked on coding. Pick Llama 3.1 70B Instruct for a larger context window (131K against 41K tokens) and work measured by LiveCodeBench (Artificial Analysis run) or Terminal-Bench Hard (Artificial Analysis run), where it leads.

Which is cheaper, Llama 3.1 70B Instruct or Qwen3 30B A3B?

Llama 3.1 70B Instruct costs $0.56 per million input tokens and $0.56 per million output tokens; Qwen3 30B A3B costs $0.12 and $0.50. Cost per solved task is known for neither of them.

Which has the bigger context window, Llama 3.1 70B Instruct or Qwen3 30B A3B?

Llama 3.1 70B Instruct, with 131K tokens against 41K.

How many benchmarks is the Llama 3.1 70B Instruct vs Qwen3 30B A3B comparison based on?

43 benchmarks have a result for both models as of October 22, 2026. Llama 3.1 70B Instruct's coding index rests on 0 independent benchmarks and Qwen3 30B A3B's on 0. Where only one model has a result, the page says so instead of naming a winner.

Compare next